Tracing your ancestry via DNA is becoming a popular pastime, thanks to a growing number of consumer tests available over the Internet. At least two-dozen companies sell tests ranging in price from $100 to $900, and public interest is thriving. Most of these tests, however, paint a very rough picture of an individual’s ancestral origins: they’re limited to the direct maternal or paternal line. But that is beginning to change.

New technologies are allowing scientists to search for markers across the genome that can more precisely predict ancestry. Much of that data is being poured into public databases, supplying much more accurate and detailed information to genetic-testing companies and new consumer tests.

The basics of genetic ancestry testing are this: scientists search for genetic markers that appear more frequently in one population than in another. By combining the information gleaned from a number of markers–anywhere from tens to thousands–researchers can estimate the percentage of an individual’s ancestry from different parent populations.

Most genetic ancestry testing to date has focused on genetic markers in the mitochondria, which everyone inherits from their mothers, and on the Y chromosome, which males inherit from their fathers. Commercial tests using these markers have sparked harsh criticism from the genomics community, which contends that the public doesn’t adequately understand the limited view of an individual’s origins that the tests provide.

Now, both the scope and resolution of genomic ancestry studies are growing dramatically, thanks to specially designed microchips that allow scientists to quickly scan hundreds of thousands of spots on an individual’s genome. That means that researchers can gather genetic information from more people in more places, generating better ancestry markers. In addition, the ability to find markers across the entire genome, rather than just within mitochondrial or Y-chromosome DNA, generates a more complete picture of one’s ancestry.

One recent effort has focused on distinguishing subsets of the European population, a challenge because, as a historically young population group, Europe has a comparatively low level of genetic variability. (Africa, on the other hand, has more genetic diversity than the rest of the world combined.) “We can easily determine the difference between northern and southern [European] populations and a number of different groups within either,” says Michael Seldin, a geneticist at the University of California, Davis. “However, in some cases, it’s difficult. There are lots of Italian individuals we can’t separate from Greeks, and Northern Italians we can’t separate from the Spanish.”

In a study published last year, Mark Shriver and his team at Pennsylvania State University analyzed 10,000 genetic markers in nearly 300 people of Armenian, Jewish, Greek, Spanish, Basque, French, Italian, German, English, Irish, Polish, and Finnish descent. They found that genetic profiles differed from north to south and east to west. “Genetic displays seem to fit with geographic features of a map,” says Shriver, who is also a consultant for DNAprint, a genetic-testing company in Sarasota, FL. For example, the Iberian Peninsula, isolated from the rest of Europe by the Pyrenees, seems to harbor a distinct genetic profile.And areas in the middle of Europe have a profile somewhere between that of the north and the south.

Two companies have already condensed those findings into commercial tests: DNAprint and 23andMe, a personal-genomics startup based in Mountain View, CA. DNAprint’s European service uses a subset of 1,349 genetic variations from the original 10,000 to classify an individual’s ancestry according to five groups: southeastern European, Iberian, Basque, continental European, and northeastern European. 23andMe, which offers ancestry analysis as part of a broader genetic screening service, estimates users’ genetic similarity to 14 different populations around the world, including northern and southern Europeans.

Consumer ancestry testing, however, remains far from exact. All genetic ancestry tests are probabilistic: while individual markers might be more likely to appear in certain populations, that is not always the case, meaning that not everyone who carries that variation has ancestors in that group.

And the profile that a particular service spits out depends on the database used to calculate it. DNAprint offers a $240 global ancestry test, AncestryByDNA 2.5, that analyzes 176 markers derived mainly from studies of four groups: Native Americans, East Asians, West Africans, and Europeans. Because those groups have contributed most heavily to the current U.S. population, the test works best for people in the States.

It may generate false results for people in other parts of the world. For example, some Europeans who took the test were deemed to have Native American ancestry. “That’s ludicrous,” says Mark Stoneking, a geneticist at the Max Planck Institute for Evolutionary Anthropology, in Germany. “It’s most likely picking up central Asian ancestry, because there is no central Asian ancestry in the databases.”